CMU-Q 15-381 Lecture 1: Introduction – AI, basic definitions, problems, road map Teacher: Gianni A. Di Caro
O UTLINE § AI? Representation and Problem Solving? § What this course is about § Fundamental definitions and notions § Road map 2
AI: I N FICTION In fiction, popular views … 3
AI: I N GAMES / REAL WORLD 2016 Libratus wins $1.7m in chips at Texas Hold’em AI system from CMU has beaten four of the world’s best poker players in a 20-day tournament. In the real world … 2016-2017 AlphaGo beats world’s top Go players AI system from DeepMind /Google CMU has beaten Go’s grandmasters in 3-games matches 4
G AMES AND AI 5
AI: I N THE REAL WORLD 2016 J. Bezos’ WP news writer, since Rio Olympics In the real world … 2014 Alexa: virtual personal assistant, smarthome 6
AI: I N THE REAL ROBOTIC WORLD In the robot world … 7
AI: I N THE DIGITAL WORLD In the daily digital life … 8
H OW DO WE LABEL X AS AN AI? Many different views, but all share a common concept … The science of making machines do things that would require intelligence if done by man (Bertram Raphael) Intelligence ? 9
I NTELLIGENCE , OPERATIONAL Artificial intelligence is that activity devoted to making machines intelligent , and intelligence is that quality that enables an entity to function appropriately and with foresight in its environment (Nils J. Nilsson, The Quest for Artificial Intelligence, 2009) ( Intelligence ) The cognitive ability of an individual (entity) to learn from experience, to reason well, to remember important information, and to (effectively) cope with the demands of daily living (Robert Sternberg) 10
A RTIFICIAL I NTELLIGENCE Artificial intelligence is that activity devoted to making machines intelligent, and intelligence is that quality that enables an entity to function appropriately and with foresight in its environment (Nils J. Nilsson) Is quicksort an example of AI? Is a pocket calculator an example of AI? q Scale q Autonomy q Speed Intelligence ~ q Flexibility Human-related notion q Generality q Capability q …. 11
T HE BIRTH OF AI (1956) § Participants included Marvin Minsky, John McCarthy, Claude Shannon, Ray Solomonoff, Arthur Samuel, Allen Newell, Herbert Simon § “ We propose that a two-month, ten-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, NH. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. ” 12
A PPROACHES TO AI The science of making machines that: Theory of the mind ⟷ Laws of thought to AI computer models, represent and reason Cognitive science about things: Logics Thought processes and reasoning Think rationally Think like people Fidelity to human performance Ideal performance 973537498401 ? Act rationally Act like people Behavior An agent that does the “right thing” based Functionally equivalent, imitation: on what it knows (the thought process it doesn’t matter how . Turing Test doesn’t matter). Rational agents 13
S TRONG VS . W EAK AI Philosopher J. Searl (1980) made a distinction between different hypothesis about AI – Weak AI: An artificial intelligence system can (only) act like it thinks and has a mind Strong AI: An artificial intelligence system can think and have a mind R&N, Ch. 26 14
N ARROW VS . G ENERAL AI – Narrow AI: An artificial intelligence system that replicates and maybe surpasses human intelligence for a dedicated purpose. “Applied” AI to a “narrow” domain. General AI: A general-purpose artificial intelligence system that replicates and maybe surpasses human intelligence . A system with comprehensive knowledge and cognitive capabilities. Processing speed and capacity can be larger than humans. Embodiment (?) → Strong AI 15
G ENERAL AI AND AI SAFETY Elon Musk: AI is “our greatest existential threat.” Stephen Hawking: “Success in creating AI would be the biggest event in human history. Unfortunately, it might also be the last...” Bill Gates: “First, the machines will do a lot of jobs for us and not be super intelligent. That should be positive if we manage it well. A few decades after that, though, the intelligence is strong enough to be a concern.” 16
T HE TECHNOLOGICAL S INGULARITY Emergence of superhuman intelligence Key idea: self-improvement Source of name: analogy between inability to predict events after the development of a superintelligence, and the space-time singularity beyond the event horizon of a black hole Some predict: this century Others argue: never 17
18
19
AI: D ATA S CIENCE + S UPERCOMPUTING § Data, data, data… to learn from! § Sensors § Internet § Social nets § Clouds § Smartphones § Fast computers § GPUs § New techniques http://www.nature.com/nature/journal/v518/n7540/abs/nature14236.html 20
T HIS COURSE : AI ~ A CT R ATIONALLY Theory of the mind ⟷ Laws of thought to AI computer models, represent and reason Cognitive science about things: Logics Thought processes and reasoning Think rationally Think like people Fidelity to human performance Ideal performance Act rationally Act like people Behavior An agent that does the “right thing” based Functionally equivalent, imitation: on what it knows (the thought process it doesn’t matter how . Turing Test doesn’t matter). Rational agents 21
R ATIONAL D ECISION -M AKING This course: “Solving problems in the best possible way!” = Computational Rationality + (Modeling + Problem Solving) = Mathematical and computational techniques for rational decision-making ↔ Act rationally § Provably optimal § Provably sub-optimal § Heuristic Fundamental question: How to issue the sequence ( ≥ 1) of decisions that best/optimally (given accessible information and data) achieve my performance objectives ? 22
D ECISION -M AKING … a single decision, or a sequence of decisions ... 23
R ATIONALITY ? Not to be absolutely certain is, I think, one of the essential things in rationality (Bertrand Russel) Definitions are the guardians of rationality, the first line of defense against the chaos of mental disintegration (Ayn Rand) In everything, one thing is impossible: rationality (Friedrich Nietzsche) 24
A GENT R ATIONALITY , O PERATIONALLY 1. Representation of the Problems / World Modeling!!! I. Newton More things should not be used than are necessary (Ockham) 2. Declare Goals and Preferences → Performance measure: Objective criterion to asses degree of success Rationality 3. Use all the necessary ≠ Omniscience available information 4. Make decisions ( act ) that, (from sensors+built-in) given (1+2+3), maximize the expected performance (with formal guarantees) 25
L EARNING O BJECTIVES 1. How to make effective abstractions in order to categorize problems and define efficient problem representations Abstraction Formal Problem Model Problem Class 2. How to define goals, preferences, and utilities to effectively direct the problem solving process I have different preferences I don’t have a specific goal, but different My goal is to states have a get to West-Bay different utility (possibly in the (numeric value) minimum time) 26
L EARNING O BJECTIVES 3. How to use all the available information to compute, or learn , optimal decisions : defining how to act Plan: (action a, action b, action c, … action n ) → goal Plan: ( 𝑦 , = 𝑏 , 𝑦 / = 𝑐 , 𝑦 1 = 𝑑 , … 𝑦 3 = 𝑜 ) → goal Open loop control Action a Policy: ( state → action) → maximize utility Closed loop control State A State B Classification: pattern → class Learning a discrete mapping Learning a function Regression: observed input → predicted output 27
(R ATIONAL ) A GENT Environment Sensors States of the Percepts representation: Agent ? feasible configurations expressed in terms of selected variables of interest Actuators Actions Agent Function, for a Rational Agent: For each possible input, the agent selects an action (1) that is expected to maximize the performance measure (2), given the evidence provided by the percept sequence (3) if need, and whatever built-in knowledge the agent has (4) § Rational: maximally achieving pre-defined goals (based on representation, accessible data and information) § Rationality only concerns what decisions are made (not the thought process behind them) 28
E XAMPLE : V ACUUM -C LEANER W ORLD § Elements: Vacuum-cleaner robot, dirt, two (interconnected) rooms § Actions: Left, Right, Suck, NoOp § World States: 8 feasible configurations 29
V ACUUM -A GENT F UNCTION § Actions: Left, Right, Suck, NoOp § Percepts: Location and Status, e.g., [A, Dirty] function Vacuum-Agent ([location, status]) returns an action • if status = Dirty then return Suck • else if location = A then return Right • else if location = B then return Left Is the agent behavior rational? 30
Recommend
More recommend
